- Logic Models
- Flow Charts
- Decisions
- Discounting
- Systems
- Linear Optimization
- Cost Benefit
- Peer Effects
- BoilerPlate
- Grading
- Intro
Learning is not what is poured from the pitcher, but what lands in the glass.
-Tara Gray and Laura Madson
Logic Models
A logic model is a representation (usually graphic) of how a program or organization produces results. It shows the relationship between what goes into a process and what comes out of a process by specifying inputs (resources, personnel, clients), activities, outputs, and outcomes and how these are connected.
Many, not most, logic models are poorly thought out and probably reflect the sorry state of the programs and organizations that produce them. But that does not mean they need to be bad and it does not mean they are useless, even if poorly constructed. As a starting point for critical analysis even a bad logic model can be useful. It gives you something concrete with which to engage when trying to figure out how some outcome is supposed to be generated by a process or organization.
Class 1 Introduction to logic models, why, what, how
Prep
- Watch animated video Logic Models for Extension (4:07)
- Logic Models in Public Health (9:56)
- W.K. Kellogg Foundation. 2004. Logic Model Development Guide (scan table of contents and introduction and read chapter 1)
- Write up problem 460, an example from the Kellogg reading.
- Work through Exercise 1 in Kellogg prior to class.
Class 2 More
Prep
- Brief introduction to program logic models (outcomes models) Paul Duignan (DOVIEW Software ) (5:03)
- Finish reading Kellogg booklet
Lab Building Logic Models in MS Office
Prep
- TBA
Skills
Describe components of logic model, interpret and criticize logic model, translate project description into conventional logic model
Problems
Wiki Pages
See also
Portfolio
Problem Set 1
Workshop 1
Workshop 2
Lab
That's the wrong way to think about it. Don't try to think about it all at once.
- Jeb Rubenfeld
Flow Charts
A flow chart is a graphic representation of the logic of a deterministic process that can contain contingencies and repetition.
Learning to draw a flow chart of an existing process or being able to read the logic of a process from a flow chart helps you to see the taken-for-granted and invisible structure of a process - the decisions that are built into it. It forces you to be explicit about what choices will be made within a process and to account for all the logically possible contingencies. Hewing to a particular discipline of flow chart drawing forces careful thinking at exactly those moments where the urge to hand-wave is strongest.
The flow chart to the left, for example, represents the logic of "keep writing research grants until you get funded and then do the research." Or, a little more pedantically: write a grant. If you get funded, do the research. Otherwise go back and repeat step one - go back to writing grants.
We will start with simple flow charts of processes with sequences and decisions and using flow charts to model classification processes. Then we move on to the representation of repetition. Then we will introduce the concepts of "black boxing," "step-wise refinement," and "deferring detail" as techniques for managing complexity. Finally, we will add time and space to flow charts to show how they can be adapted for project management tasks.
Class 1 Prerequisites, if/then/else, basics, action, decision, loops
Prep
- Ryan: Flow Charts, An Introduction
- Stoked & Zeckhauser, pp. 9-10
- Problems 71, 72
- Univ Plymouth, UK: Flow Charts for Simple Tasks: Tutorial with exercises
- Univ Plymouth, UK: Flow Charts for Classification: Tutorial with exercises
- (optional) Wikipedia Flow Charts
Class 2 Stepwise refinement, division of labor
Prep
- Ryan: Stepwise Refinement
- Ryan: Division of Labor
- Wirth, N. 1971. "Program Development by Stepwise Refinement." Communications of the ACM, Vol. 14, No. 4, April 1971, pp. 221-227. (OPTIONAL)
- Problems 91, 244, 322, 323, 248, 249
Lab
Prep
- About.COM: the IF Function)
- MS Excel Help: Switch between relative, absolute, and mixed references
- Problems 73, 88, 89, 90
Skills
Create a flow chart from a textual description.
Use flowchart to demonstrate stepwise refinement and black-boxing
Create and interpret flow chart that includes time and divisions of labor
Offer critique and corrections to a poorly formed flowchart
Draw flow charts in Microsoft Excel, Word or other software
Identify and comment on concerns and caveats. What are flow charts good for, what not?
All Problems
Wiki Pages
See also
Stoicism Flowchart
An overview by HCI consulting in Australia
Gehani, N. 1981. "Program Development by Stepwise Refinement and Related Topics." Bell System Technical Journal, vol. 60, no. 3.
Wirth, N. 1971. "Program Development by Stepwise Refinement." Communications of the ACM, Vol. 14, No. 4, April 1971, pp. 221-227.
Decision Trees
A decision tree is basically a non-deterministic flow chart and so we move on to them after we do flow charts. This module starts with a review of basic probability. Then we look at simple decisions under uncertainty, perfect tests and the value of information, imperfect tests and "tree flipping," using decision trees to account for risk and non-monetary outcomes. In labs we will enhance our drawing skills and learn how to use "controls" to make a model interactive and dynamic.
This module expects you to have a good grasp of basic probability. The concept of "expected (monetary) value" (emv) is used extensively in decision tree analysis. You have likely encountered this before, but if not, review it now on this worksheet.
Class 1 Math pre-reqs; Basic tree analysis, folding back (dt01Probability dt02Expected Value dt03Translation dt04Simple tree analysis)
Prep
- Ryan: Math refresher and Probability Review
- Stokey & Zeckhauser, ch. 12, pp. 201-215
- Ryan: "Decision Analysis I"
- Problems 2, 3, 92, 93
- Wikipedia on Decision Analysis
Class 2 The value of information (dt05)
- Stokey & Zeckhauser, ch. 12, pp. 215-221
- Ryan: Working Backward
- Ryan: The Value of Information
- Problems 95, 100
Lab: Building a Decision Model in Excel (part 1)
Prep
- Read about TEXT(), conditional formatting, spinner controls in Excel help
Class 3 Imperfect tests (dt07) (slides, worksheet (pdf, docx))
Prep
- Stokey & Zeckhauser, ch. 12], "Imperfect Tests," pp. 221-236
- Ryan: Tree flipping and Imperfect Tests
- Problems 102, 103
Class 4 Risk and Utility (dt06, dt07) (Slides: PPTX,PDF)
Prep
- Stokey & Zeckhauser, ch. 12, "Allowing for Risk Aversion," pp. 216-219
- Ryan: Including Risk
- Stokey & Zeckhauser, ch. 12, "Utility Theory," pp. 236-254
- Problems tba
Lab: Building a Decision Model in Excel (part 2)
Prep
- tba
Skills
- outcomes & events
- mutually exclusive & exhaustive
- independence
- P(A or B)
- P(A and B)
- P(A|B)
- $P(x) \times \mbox{value of outcome X}$
- Expected value of a set of mutually exclusive alternatives (e.g., branches coming out of a chance node) is sum of expected values.
- Read a description, draw the tree
- See a tree, write the description
- Work backwards to calculate values of chance nodes and pick decisions at choice nodes.
Use decision tree to calculate the value of information provided by a perfect test (how much should we be willing to pay for information?)
- Take an ordinary DT and add a test/no test choice node, calculate value of test
- Expected value of perfect information (EVPI)
- Define false positive and false negative
- "Flip" a tree to calculate expectations of test outcomes
Explain/distinguish related concepts: choice/chance nodes, exhaustive/mutually exclusive, folding-back, imperfect tests, risk-aversion, the value of information, utility theory
When are DTs useful? When should they not be used? Identify and respond to typical concerns and caveats
Problems
Wiki Pages
See also
Files
A sparrow in thy hand is better than a thousand sparrows flying.
- Ahiqar, 6th century BCE
Discounting
Why exactly is a bird in the hand worth two in the bush? Is this actually true? In this module we will learn how to compute the present value of a stream of costs or revenues so that comparisons can be made between different scenarios in which costs and benefits occur over a span of time.
Class 1 Basics Slides, Handout
Prep
- Math Prereqs
- Stokey & Zeckhauser Ch. 10, "The Valuation of Future Consequences: Discounting," pp. 159-165
- Ryan: "Discounting"
- Problems TBA
Class 2 internal rate of return, payback; Where do discount rates come from? Caveats and issues
Prep
- Stokey & Zeckhauser Ch. 10, "The Valuation of Future Consequences: Discounting," pp. 165-176
- Problems TBA
Lab NPV and its friends
Prep
- TBA
Skills
Explain in plain language what discounting is and why it is important.
Demonstrate dexterity in working with compound interest formula.
Given a schedule of costs and benefits and a discount rate, calculate NPV both "by hand" using use Excel with/without NPV() related functions
Explain where discount rates come from in practice and identify rates currently used in several well known contexts
Qualitatively interpret generic discount rate/PV curves
Explain/distinguish related concepts: present value, discount rate, internal rate of return, break even and payback periods, opportunity cost
Identify and respond to typical concerns and caveats around discount rates.
Problems
See also
Portfolio
Problems:
"The overall name of these interrelated structures is system. The motorcycle is a system. A real system. …There's so much talk about the system. And so little understanding. That's all a motorcycle is, a system of concepts worked out in steel. There's no part in it, no shape in it that is not in someone's mind. I've noticed that people who have never worked with steel have trouble seeing this- that the motorcycle is primarily a mental phenomenon."
- Robert Pirsig Zen and the Art of Motorcycle Maintenance
Feedback and Flow
Up to now we have modeled the decisions of individual agents. Now we will transition to thinking about modeling the behavior of systems as a whole. The system may be composed of many agents, each responding to environmental conditions which may be effects their behavior (that is, they may be subject to feedback). We will start with difference equations and then slide into stock and flow models and systems dynamics.
Class 1 Math prereqs; system behaviors; difference equations Slides (PDF,PPTX, KEY)
Prep
- Review Using Subscripts if mathematical subscripts are not familiar
- Stokey & Zeckhauser, pp. 47-58
- Ryan: Difference Equations, An Introduction
- Problems: 107 | 108 | 109 | 110 | 111 | 5
Class 2 Equilibria
Prep
- Stokey & Zeckhauser, pp. 58-66
- Ryan: "Rates and Amounts"
- Ryan: "Equilibria"
- Problems: 107 | 108 | 109 | 110 | 111 | 5 (WITH full charts, etc.)
Lab Difference Equations
Prep
- TBA
Class 3 From difference equations to stock, flow, and feedback (Slides: pdf | pptx | key)
Prep
- Stokey & Zeckhauser, pp. 66-73
- Ryan: Stock and Flow Models
- Ryan: Stock and Flow Models II
- Wikipedia Stock and Flow and Systems Dynamics Models
- Kirkwood, Ch. 1 "System Behavior and Causal Loops"
- Problems: 132,134,135
Class 4 Systems dynamics simulation (Slides: PDF, PPTX, KEY)
Prep
- Kirkwood, Ch. 2 "A Modeling Approach"
- Problems: TBA
Lab Stock and Flow Models
Prep
- Stokey & Zeckhauser, pp. 89-97
Skills
Be comfortable using subscripts and understanding what we mean when we write $X_{n} = 2 \times X_{n-1}$ and be able to work out the value of, say, $X_{4}$ given $X_{0}$ and an expression for $X_{n}$.
Translate a problem into difference equations
Recognize in a problem what the quantities of interest are, what the time periods are, and what amounts and rates represent the differences from one period to the next.
Create and Use Pn+1 vs. Pn charts to assess difference equation system behavior
- Constructing difference equation chart from equations/table with 45 degree line
- Be able to explain how chart is made and what a point (x,y) on the line means, explain what 45 degree line represents
- Based on slope of DE line, describe kinds of equilibria, if any
- Describe difference equation in Pn+1 = a Pn + b terms and calculate Pe from this.
Implement difference equations model in Excel and graphically display outcome"]]
- Translate standard subscript notation into Excel formulas
- Use autofill to extend integer time periods and formula for quantity of interest
- Use other cells to contain parameters that are referred to in the difference equation formula
- Make appropriate use of relative and absolute cell references so that formulas autofill correctly
- Use a "spinner" to make a parameter variable; use a "divide by (a power of) ten" formula to make a spinner control parameters of arbitrary precision.
- Format text in Excel to show subscripts.
- Create XY scatterplot charts with legends and titles depending on data
Convey understanding of relationship between difference equations and related models/techniques.
Describe how to get from difference equations to stock and flow models
Suggest indications and contraindications, identify and respond to typical concerns and caveats
Describe assumptions and caveats
stock
flow
valve
information
positive/negative feedback
causal loops
equilibrium
Translate system description into causal loop diagram
Translate system description into stock and flow diagram
Translate stock and flow diagram into difference equations
Describe generic system performance possibilities (oscillating, growing, asymptotic, damped, etc.)
Find a source for system dynamic software
Stock and flow models are a part of an approach to modeling called systems dynamics. A search on that term will get you all you need.
Caveats
Problems
Wiki Pages
See also
If you optimize everything, you will always be unhappy.
- Donald Knuth
Linear Programing and Optimization
With decision trees we learned how we can make the best choice among alternatives. But what about when we need to choose a best combination of things to accomplish our objectives?
Class 1 Math prereqs, basic concepts, graphing (Slides: PDF, PPTX, KEY, class handout)
Prep
- Stokey & Zeckhauser, Ch.11, "Linear Programming" 177-188
- Ryan: Math Prereqs for LP
- Ryan: Linear Programming I
- Problems 219 220 221 223 224
Class 2 Shadow Prices and Sensitivity Analysis
- Stokey &Zeckhauser, Ch.11, "Linear Programming," pp. 188-200
- Excel Easy: Solver Tutorial
- Problems 327 226 227 228 229 230 231
Lab Using Solver with Excel
- Excel Easy: Solver Tutorial
Class 3 Intermediate Considerations
Prep
- TBA
- Problems tba
Class 4 Issues and Caveats
Prep
Lab More with Solver
Skills
Understand vocabulary and concepts associated with linear programming.
Translate word problems into inequalities for use in linear programming model.
Be able to solve linear programming model graphically
Translate simple linear programming problem into Excel and use solver to find optimum.
Suggest indications and contraindications, identify and respond to typical concerns and caveats
Problems
See also
[The Freedom of Information Act is] the Taj Mahal of the Doctrine of Unanticipated Consequences, the Sistine Chapel of Cost-Benefit Analysis Ignored.”
- Antonin Scalia
Cost Benefit Analysis
Preamble. Integer ut lobortis lorem, in tristique velit. Curabitur ante metus, mollis vitae dolor quis, pellentesque semper enim. Nulla nec fringilla nibh, posuere gravida erat. Nunc sed odio luctus, efficitur enim venenatis, suscipit ante. In ac orci blandit, iaculis ligula at, aliquet tortor.
Class 1 Concepts, Fundamental Rule, The Four Cases
Prep
- Stokey & Zeckhauser, ch. 9, pp. 134-45
- Ryan: Cost Benefit Analysis
- Wikipedia Cost Benefit Analysis
- Problems TBA (on concept and recognizing types)
Class 2
- Stokey & Zeckhauser, ch. 9, pp. 146-58
- Problem set 1
Lab
Prep
- TBA
Class 3
Prep
- TBA
Class 4 Issues and Caveats
Prep
- TBA
Lab
Prep
- TBA
Skills
Appreciate and explain basic idea of (global/aggregate) net benefit and potential compensation
Calculating net cost/benefit and marginal cost/benefit. Sketch charts of same. Interpret tables/charts.
Identify four choice scenarios and the CB protocol appropriate to each.
Produce and interpret flow charts describing these
Explain/distinguish related concepts: cost, benefit, net, marginal, benefit/cost ratio, willingness-to-pay, compensation, cost-effectiveness
Identify and respond to typical concerns (especially separation of CB and distributional issues) and caveats
Problems
Wiki Pages
See also
Portfolio
Problem Set 1
Workshop 1
Workshop 2
Lab 1
Workshop 3
Workshop 4
Lab 2
Peer Effects
# Aggregation, Sorting, and Peer Effect Models
# Schelling, Conway, Granovetter
# Tipping Points, Diffusion, and Contagion
# Actors, Games, Coordination, and Cooperation
Problems
See also
Portfolio
Attendance As a graduate class, 100% attendance is expected. You are responsible for obtaining from classmates or other sources any materials missed because of absences. Do not contact the instructor with valid excuses. Attendance at lab, in particular, is expected to be 100%; missed labs may result in final grade attenuation at instructor's discretion.
Class Preparation and Assignments You are expected to read, work with, and learn from assigned readings BEFORE the class in which they will be discussed. Do not expect lectures and notes you might take during them to suffice for learning the material. Written assignments are due when they are due without exception. Expect zero credit and zero feedback on any work not submitted by deadlines. Better incorrect and incomplete but on time. Incorrect or incomplete work should still be presented in as professional a manner as possible.
Digital Tools: This course makes extensive use of various digital tools. You are expected to check your email at least once every 24 hours for potential course updates. Students will maintain a digital portfolio and most work will be submitted digitally. Please inform instructor if lack of access to hardware or network poses an impediment to course work. Syllabus is subject to adjustment based on class members' progress; current syllabus always online at http://ppol225.danryan.us.
Academic Integrity In the academy, plagiarism is a serious breach of trust that destroys one's credibility. Academic dishonesty will not be tolerated in this course. Academic dishonesty includes all types of cheating, misrepresentation of authorship, failure to properly cite sources, as well as deliberate attempts to plagiarize (that is, intentionally or knowingly using someone else’s ideas, words, and/or thoughts without acknowledging the source). All work for which another source is not cited is assumed to be that of the writer. ALL material taken from another source must be cited by in-line references with appropriate information contained in a bibliography or list of works cited. Material from field notes or interviews must be referenced as such.
AccessibilityTo request academic accommodations due to a disability, students should contact Services for Students with Disabilities in the Cowell Building. If you have a letter indicating you have a disability which requires academic accommodations, please present the letter to me so that I will be able to provide the accommodations that you need in this class.
Grading: Grades based on class participation (= doing problems and other portfolio work during the course of the semester) (33.3%); module exams (33.3%); final exam (33.3%).
Grading scheme translation:
A = demonstrated excellence
B = demonstrated competence
C = suggested misunderstanding
D = demonstrated misunderstanding
F = not attempted
Course Intro
What is a modeling? What is simulation? Why learn how to do them? What does the course look like? How will we proceed?
To Do List
- Obtain books.
- Ascertain web access and familiarity with digital tools.
- Attend and participate in class and lab.
- Pre-Test.
Agenda
Ice Breaker Visual Telephone
User's Guide to Digital Tools
Study Hall read pp. 13-17 of Stokey & Zeckhauser
Workshop: Types of Models
Lab: Lab 1 Basic Digital Skills
Problems
See also
Portfolio
Lab demo document
Zotero Account
Share a folder